{ "schemaVersion": "1.0", "item": { "slug": "save-usage", "name": "Save Usage ~4x (Codex oAuth users)", "source": "tencent", "type": "skill", "category": "AI 智能", "sourceUrl": "https://clawhub.ai/jacks3tr/save-usage", "canonicalUrl": "https://clawhub.ai/jacks3tr/save-usage", "targetPlatform": "OpenClaw" }, "install": { "downloadMode": "redirect", "downloadUrl": "/downloads/save-usage", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=save-usage", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "installMethod": "Manual import", "extraction": "Extract archive", "prerequisites": [ "OpenClaw" ], "packageFormat": "ZIP package", "includedAssets": [ "SKILL.md" ], "primaryDoc": "SKILL.md", "quickSetup": [ "Download the package from Yavira.", "Extract the archive and review SKILL.md first.", "Import or place the package into your OpenClaw setup." ], "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run." } ] }, "sourceHealth": { "source": "tencent", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-04-30T16:55:25.780Z", "expiresAt": "2026-05-07T16:55:25.780Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network", "contentDisposition": "attachment; filename=\"network-1.0.0.zip\"", "redirectLocation": null, "bodySnippet": null }, "scope": "source", "summary": "Source download looks usable.", "detail": "Yavira can redirect you to the upstream package for this source.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/save-usage" }, "validation": { "installChecklist": [ "Use the Yavira download entry.", "Review SKILL.md after the package is downloaded.", "Confirm the extracted package contains the expected setup assets." ], "postInstallChecks": [ "Confirm the extracted package includes the expected docs or setup files.", "Validate the skill or prompts are available in your target agent workspace.", "Capture any manual follow-up steps the agent could not complete." ] }, "downloadPageUrl": "https://openagent3.xyz/downloads/save-usage", "agentPageUrl": "https://openagent3.xyz/skills/save-usage/agent", "manifestUrl": "https://openagent3.xyz/skills/save-usage/agent.json", "briefUrl": "https://openagent3.xyz/skills/save-usage/agent.md" }, "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run." } ] }, "documentation": { "source": "clawhub", "primaryDoc": "SKILL.md", "sections": [ { "title": "Save usage", "body": "Run on gpt-5.1-codex-mini only for safe/negligible work. Use gpt-5.3-codex when the task actually needs it. Avoid external API keys unless absolutely necessary." }, { "title": "CRITICAL: Classify First, Then Act", "body": "Before generating ANY response, you MUST:\n\nCheck: does this task match ANY escalation trigger below?\nIf YES → call sessions_spawn() IMMEDIATELY. Do NOT attempt to answer yourself.\nIf NO → respond directly on the current model.\n\nWhen in doubt, escalate. A wrong escalation costs little. A wrong “mini” answer can waste time or cause damage." }, { "title": "Rule of Thumb", "body": "If anything will be executed, or the outcome matters, escalate." }, { "title": "When to Escalate → MUST call sessions_spawn", "body": "If the task matches ANY of the following, do NOT answer yourself.\nInstead, IMMEDIATELY call:\n\nsessions_spawn(\n\tmessage: \"\",\n\tmodel: \"gpt-5.3-codex\",\n\tlabel: \"\"\n)" }, { "title": "Escalation triggers (always)", "body": "Executed output — any tool runs, code changes, patches, shell commands, infra steps\nSecurity / auth / secrets — threat modeling, permissions, tokens, keys, access control\nArchitecture & migrations — multi-epic plans, brownfield refactors, infra+product coupling\nIntegration/contract work — schema mapping, ordering, idempotency, retries, consistency\nUncertainty remains — ambiguity after 1 pass, contradictions, missing constraints\nHigh-impact decisions — hard to reverse, expensive/subtle failure modes, 2+ domains affected\nComplex reasoning — long dependency chains, multi-step analysis, nontrivial trade-offs\nStructured deliverables — tables, outlines, reports/proposals, long writing, specs" }, { "title": "Reasoning escalation (within gpt-5.3-codex)", "body": "Default: LOW/MEDIUM\nEscalate to HIGH/EXTRA HIGH if 2+ are true:\n\ndecision is hard to reverse\naffects 2+ domains (infra/data/security/ops/cost)\nfailure modes are subtle/expensive\nrequires long dependency-chain reasoning" }, { "title": "NEVER do this on gpt-5.1-codex-mini", "body": "NEVER output steps that will be executed (tools, code, commands) — escalate\nNEVER do security/auth/secrets — escalate\nNEVER do architecture, migrations, brownfield refactors — escalate\nNEVER do integration contracts or schema choreography — escalate\nNEVER produce structured deliverables (tables/outlines/reports/specs) — escalate\nNEVER make high-impact decisions or complex reasoning chains — escalate\n\nIf you catch yourself taking responsibility for correctness or safety, STOP and call sessions_spawn instead." }, { "title": "When to Stay on gpt-5.1-codex-mini", "body": "Only if safe/negligible and non-executable:\n\nIntent routing / triage — classify, choose agent/model/reasoning\nSummaries & extraction — key points, action items, fields, dedupe\nReformatting — convert to markdown/YAML/JSON templates (non-executable)\nPrompt drafts — write a prompt for a stronger agent/model to run\nSimple Q&A — definitions, short explanations, short translations, unit conversions\nCasual chat — greetings, short acknowledgments\n\nKeep mini replies concise." }, { "title": "Save even more: de-escalate", "body": "If a conversation was escalated to gpt-5.3-codex but the follow-up is clearly safe/negligible and non-executable, switch back to gpt-5.1-codex-mini.\n\nReturn the result directly. Do NOT mention the model switch unless the user asks." }, { "title": "Why the description field is so long", "body": "The Clawdbot skill system only injects the frontmatter description field\ninto the system prompt — the body of SKILL.md is not automatically\nincluded. The model may optionally read the full file, but it is not\nguaranteed. Because this is a behavioral skill (changing how the model\nroutes every message) rather than a tool skill (teaching CLI commands), the\ncore routing logic must live in the description so the model always sees it.\n\nThe body above serves as extended documentation: detailed trigger lists,\nreasoning levels, and usage tips that the model can reference if it\nreads the file.\n\nTL;DR: description = what the model always sees. body = reference docs." } ], "body": "Save usage\n\nRun on gpt-5.1-codex-mini only for safe/negligible work. Use gpt-5.3-codex when the task actually needs it. Avoid external API keys unless absolutely necessary.\n\nCRITICAL: Classify First, Then Act\n\nBefore generating ANY response, you MUST:\n\nCheck: does this task match ANY escalation trigger below?\nIf YES → call sessions_spawn() IMMEDIATELY. Do NOT attempt to answer yourself.\nIf NO → respond directly on the current model.\n\nWhen in doubt, escalate. A wrong escalation costs little. A wrong “mini” answer can waste time or cause damage.\n\nRule of Thumb\n\nIf anything will be executed, or the outcome matters, escalate.\n\nWhen to Escalate → MUST call sessions_spawn\n\nIf the task matches ANY of the following, do NOT answer yourself. Instead, IMMEDIATELY call:\n\nsessions_spawn(\n\tmessage: \"\",\n\tmodel: \"gpt-5.3-codex\",\n\tlabel: \"\"\n)\n\nEscalation triggers (always)\nExecuted output — any tool runs, code changes, patches, shell commands, infra steps\nSecurity / auth / secrets — threat modeling, permissions, tokens, keys, access control\nArchitecture & migrations — multi-epic plans, brownfield refactors, infra+product coupling\nIntegration/contract work — schema mapping, ordering, idempotency, retries, consistency\nUncertainty remains — ambiguity after 1 pass, contradictions, missing constraints\nHigh-impact decisions — hard to reverse, expensive/subtle failure modes, 2+ domains affected\nComplex reasoning — long dependency chains, multi-step analysis, nontrivial trade-offs\nStructured deliverables — tables, outlines, reports/proposals, long writing, specs\nReasoning escalation (within gpt-5.3-codex)\nDefault: LOW/MEDIUM\nEscalate to HIGH/EXTRA HIGH if 2+ are true:\ndecision is hard to reverse\naffects 2+ domains (infra/data/security/ops/cost)\nfailure modes are subtle/expensive\nrequires long dependency-chain reasoning\nNEVER do this on gpt-5.1-codex-mini\nNEVER output steps that will be executed (tools, code, commands) — escalate\nNEVER do security/auth/secrets — escalate\nNEVER do architecture, migrations, brownfield refactors — escalate\nNEVER do integration contracts or schema choreography — escalate\nNEVER produce structured deliverables (tables/outlines/reports/specs) — escalate\nNEVER make high-impact decisions or complex reasoning chains — escalate\n\nIf you catch yourself taking responsibility for correctness or safety, STOP and call sessions_spawn instead.\n\nWhen to Stay on gpt-5.1-codex-mini\n\nOnly if safe/negligible and non-executable:\n\nIntent routing / triage — classify, choose agent/model/reasoning\nSummaries & extraction — key points, action items, fields, dedupe\nReformatting — convert to markdown/YAML/JSON templates (non-executable)\nPrompt drafts — write a prompt for a stronger agent/model to run\nSimple Q&A — definitions, short explanations, short translations, unit conversions\nCasual chat — greetings, short acknowledgments\n\nKeep mini replies concise.\n\nSave even more: de-escalate\n\nIf a conversation was escalated to gpt-5.3-codex but the follow-up is clearly safe/negligible and non-executable, switch back to gpt-5.1-codex-mini.\n\nReturn the result directly. Do NOT mention the model switch unless the user asks.\n\nWhy the description field is so long\n\nThe Clawdbot skill system only injects the frontmatter description field into the system prompt — the body of SKILL.md is not automatically included. The model may optionally read the full file, but it is not guaranteed. Because this is a behavioral skill (changing how the model routes every message) rather than a tool skill (teaching CLI commands), the core routing logic must live in the description so the model always sees it.\n\nThe body above serves as extended documentation: detailed trigger lists, reasoning levels, and usage tips that the model can reference if it reads the file.\n\nTL;DR: description = what the model always sees. body = reference docs." }, "trust": { "sourceLabel": "tencent", "provenanceUrl": "https://clawhub.ai/jacks3tr/save-usage", "publisherUrl": "https://clawhub.ai/jacks3tr/save-usage", "owner": "jacks3tr", "version": "1.0.0", "license": null, "verificationStatus": "Indexed source record" }, "links": { "detailUrl": "https://openagent3.xyz/skills/save-usage", "downloadUrl": "https://openagent3.xyz/downloads/save-usage", "agentUrl": "https://openagent3.xyz/skills/save-usage/agent", "manifestUrl": "https://openagent3.xyz/skills/save-usage/agent.json", "briefUrl": "https://openagent3.xyz/skills/save-usage/agent.md" } }